Although distinction of rhabdomyosarcoma from the other round cell tumors of the soft tissues in children is not difficult in most cases, the argument as to what connotes a primitive rhabdomyosarcoma and an extraskeletal Ewing's sarcoma is a valid one. Furthermore, this argument comprises more than a philosophical question, since patients with rhabdomyosarcoma are treated in a different protocol than those with Ewing's sarcoma. We have recently studied immunohistochemically the presence of the product of the pseudoautosomal MIC2 gene in a series of childhood tumors and found preferential expression of this product by Ewing's sarcoma and lack thereof in rhabdomyosarcoma. However, a few primitive tumors expressing 1 or 2 muscle markers (i.e. skeletal actin and/or desmin) and ultrastructural appearance favoring a primitive rhabdomyosarcoma, were found positive for the MIC2 product, while others with similar characteristics were negative. Since several muscle determination genes have recently been cloned and cDNA probes have become commercially available, we decided to study the expression of these genes in the same cases that have been studied by the MIC2 gene product and find out if these markers are more reliable of myogenesis than the MIC2 gene product or the existing antibodies against muscle proteins. For this study, DNA will be extracted from 50 micron sections of paraffin-embedded tissues and will be analyzed by Southern blotting for the presence of MyoD1, Myogenin, Myf5, Myf6, skeletal myosin heavy chain and muscle creatine kinase. The value of these markers over the more traditional ones will be critically evaluated, especially in those primitive cases that are difficult to distinguish because of conflicting immunohistochemical data.